Navegando por Assunto "Clustering"

Agora exibindo 1 - 5 de 5

Acesso aberto (Open Access)
Agrupamento de fornos de redução de alumínio utilizando os algoritmos Affinity Propagation, Mapa auto–organizável de Kohonen (som), Fuzzy C–Means e K–Means
(Universidade Federal do Pará, 2017-10-11) LIMA, Flávia Ayana Nascimento de; CARDOSO, Diego Lisboa; http://lattes.cnpq.br/0507944343674734; OLIVEIRA, Roberto Célio Limão de; http://lattes.cnpq.br/4497607460894318
The continuous development of technology accounts for measures that provide industries benefits to grant them profitability and competitive advantage. In the mineralogy field, aluminum smelting usually requires substantial number of cells, also known as reduction pots, to produce aluminum in a continuous and complex process. Analytical monitoring is essential for those industries’ competitive advantage, given that during operation some cells show behavior similar to others, thereby forming clusters of cells. These clusters depend on data patterns usually implicit or invisible for the operation, but can be found by data analysis techniques. In this work four clustering techniques are presented to that end: the Affinity Propagation; the Kohonen Self Organizing Map; the Fuzzy C–Means; and the K–Means Algorithm. These techniques are used to find and group cells that share similar behavior, by analysing seven variables which are closely related to the aluminum reduction process. This work aims at addressing the benefits of clustering, especially by simplifying the aluminum potline analysis, once a large group of cells might be summarized in one sole group, what can provide more compact yet rich information for data driven modeling and control. Moreover, the identification of similar data patterns in clusters makes the task of those who is going to be in charge of analyzing these dats. This work also identifies the ideal cluster size for each technique applied.
Acesso aberto (Open Access)
Ciência de dados e aprendizado de máquina aplicados ao estudo de variáveis epidemiológica hanseníase na Amazônia
(Universidade Federal do Pará, 2024-12-18) FALCÃO, Igor Wenner Silva; CARDOSO, Diego Lisboa; http://lattes.cnpq.br/0507944343674734; SERUFFO, Marcos César da Rocha; http://lattes.cnpq.br/3794198610723464
Leprosy is a significant public health problem that largely affects low-income populations. Although the World Health Organization (WHO) establishes guidelines for diagnosis, prevention, and treatment, disease detection faces limitations, often resulting in late or inaccurate diagnoses and leading to serious neurological complications and multidrug-resistant cases. Therefore, early diagnosis is essential to reduce the burden of this disease. Machine learning has been widely used in several areas of science and industry, but especially in health, where it plays an essential role in the analysis and treatment of large volumes of data. In this sense, this thesis investigates the application of a model based on Data Science and Machine Learning to act in the specification of the clinical profile of possible leprosy cases in the Amazon Region and, thus, to be able to act preventively in the early diagnosis and treatment of patients under medical followup. The work takes into account clinical data of patients from a non-public dataset, collected between 2015 and 2020 in the North region of Brazil. Therefore, this thesis proposes a learning model to identify groups clinically affected by the disease using Clustering and Random Forest techniques. In the results obtained, the proposed model demonstrated efficiency in evaluating the probability of individuals being ill, achieving an accuracy of 90.39% in the performance evaluation and identifying a probability of 83.46% of an individual being ill, considering a set of epidemiological and non-generic variables. This approach offers a promising vision for the future of health, allowing the formulation of effective strategies for the early identification of possible cases.
Acesso aberto (Open Access)
Clusterização de padrões espaço-temporais de precipitação na Amazônia via deep convolutional autoencoder
(Universidade Federal do Pará, 2023-07-07) SILVA, Vander Augusto Oliveira da; TEIXEIRA, Raphael Barros; http://lattes.cnpq.br/4902824086591521; https://orcid.org/0000-0003-2993-802X
Studies using different machine learning methods for knowledge discovery and pattern recognition in precipitation time series are increasingly frequent in the literature. Identify and analyze patterns in precipitation time series in a particular region is fundamental for its socioeconomic development. Therefore, it can be stated that knowledge and understanding of the rainfall characteristics of the regions are important to enable the planning of the use, management and conservation of water resources. The natural phenomenon of precipitation is a fundamental process with a direct impact on watersheds and on human and environmental development. The variability of this phenomenon has important implications for the navigability of rivers, individual abundance and species richness. In recent years, many studies with this approach have been carried out in Brazil, mainly in the Amazon region. This research aimed to develop a computational method for analyzing time series of precipitation using machine learning techniques with unsupervised learning, in order to propose an method capable of extracting complex features from the data, obtaining a map of attributes at low dimensionality for pattern recognition, discovery of homogeneous regions with respect to precipitation and approximate reconstruction of precipitation time series in the Legal Amazon. The proposed deep learning neural network model is trained to learn the main and most complex features of the original data and present them in low dimensionality in latent space. After the training, the results are promising, the observations of the reconstructed data showed a good performance as evaluated by the RMSE and NRMSE metric with resulting values equal to 0.06610 and 0.3355 respectively. The analysis of the representation of the data in low dimension was applied and analyzed by a clustering structure using hierarchical agglomerative with Ward’s method. This methodology also showed good results, as it carried out consistent groupings characterizing ho- mogeneous regions in relation to precipitation data. Thus, demonstrating that the representation in low dimensionality carried the main characteristics of the time series of the analyzed data. It is noteworthy that the method developed in this study can be applied not only in the Amazon region, but also in other areas with similar challenges related to time series analysis.
Acesso aberto (Open Access)
. Clusterização, classificação e predição de “pré-efeito anódico” de cuba eletrolítica de alumínio primário
(Universidade Federal do Pará, 2020-08-21) CONTE, Bruno Nicolau Magalhães de Souza; OLIVEIRA, Roberto Célio Limão de; http://lattes.cnpq.br/4497607460894318
The industrial sector is one of the main responsible for the serious environmental situation on the planet and also for increasing legal requirements, in relation to the waste generated. On the other hand, many companies have reacted proactively, based on the implementation of management strategies such as: clean production, environmental certification, reduction of toxic waste, recycling, sustainable consumption and reuse, mainly. It is worth mentioning that the aluminum reduction process is responsible for a large amount of greenhouse gas emissions and, thus, promotes environmental impacts and serious climate changes. During the aluminum reduction process, the occurrence of the anodic effect causes an extreme increase in the tub tension and, consequently, an increase in the bath temperature, with very high temperatures, resulting in a thermal disturbance, with the possibility of melting the insulating layer of the vat and the final consequences are the loss of production in the entire vat line, its shortened service life and the production of PFC gases. Seeking a strategy based on sustainability, I try to take into account the problem of the worsening of the Greenhouse Effect, the extreme increase in kiln tension and, consequently, the loss of production in the entire line of vats, this work proposes the use of an Artificial Neural Network together with Clustering algorithms to automatically create anodic pre-Effect labels, and thus predict the nonlinear dynamic behavior of the primary aluminum reduction industry oven anodic pre-effect, based on actual vat data electrolytic. With the use of these Machine Learning models, it is possible to predict the occurrence of the anodic pre-effect, allowing process operators to take mitigating measures to suppress the anodic effect, avoiding the loss of aluminum production in the vat and decreasing the emission of gases from the greenhouse effect.
Acesso aberto (Open Access)
Machine learning algorithms for damage detection in structures under changing normal conditions
(Universidade Federal do Pará, 2017-01-31) SILVA, Moisés Felipe Mello da; SALES JÚNIOR, Claudomiro de Souza de; http://lattes.cnpq.br/4742268936279649; COSTA, João Crisóstomo Weyl Albuquerque; http://lattes.cnpq.br/9622051867672434
Engineering structures have played an important role into societies across the years. A suitable management of such structures requires automated structural health monitoring (SHM) approaches to derive the actual condition of the system. Unfortunately, normal variations in structure dynamics, caused by operational and environmental conditions, can mask the existence of damage. In SHM, data normalization is referred as the process of filtering normal effects to provide a proper evaluation of structural health condition. In this context, the approaches based on principal component analysis and clustering have been successfully employed to model the normal condition, even when severe effects of varying factors impose difficulties to the damage detection. However, these traditional approaches imposes serious limitations to deployment in real-world monitoring campaigns, mainly due to the constraints related to data distribution and model parameters, as well as data normalization problems. This work aims to apply deep neural networks and propose a novel agglomerative cluster-based approach for data normalization and damage detection in an effort to overcome the limitations imposed by traditional methods. Regarding deep networks, the employment of new training algorithms provide models with high generalization capabilities, able to learn, at same time, linear and nonlinear influences. On the other hand, the novel cluster-based approach does not require any input parameter, as well as none data distribution assumptions are made, allowing its enforcement on a wide range of applications. The superiority of the proposed approaches over state-of-the-art ones is attested on standard data sets from monitoring systems installed on two bridges: the Z-24 Bridge and the Tamar Bridge. Both techniques revealed to have better data normalization and classification performance than the alternative ones in terms of false-positive and false-negative indications of damage, suggesting their applicability for real-world structural health monitoring scenarios.